In this work, we investigate in detail the capabilities of present (H.E.S.S., MAGIC, VERITAS) and planned (CTA) ground-based Cherenkov telescope systems to detect angular anisotropies in the diffuse gamma-ray background. We first study the impact of instrumental characteristics (effective area, field of view, angular resolution, and background rejection efficiency) to the ability to detect anisotropies. In addition, we compare different observation strategies, i.e., whether a single deep observation or a splitting over multiple shallow fields is preferred. Secondly, the sensitivity to anisotropies generated by self-annihilating dark matter is investigated for different common dark matter models. With planned configurations of CTA, we find that a relative contribution of ~10% from dark matter annihilation to the diffuse gamma-ray background can be detected, together with the sensitivity to the self-annihilation cross section = 3 10^(-26) cm3s-1 expected from thermal dark matter freeze-out. We also stress the importance of constraining anisotropies from unresolved astrophysical sources already with the current generation of instruments, as a novel and complementary method to constrain the properties of TeV sources.